Stephen Toub [Tue, 27 Nov 2018 14:27:06 +0000 (09:27 -0500)]
Use local array in ConcurrentQueue<T> for small perf improvement (#21192)
Enqueue/TryDequeue/TryPeek are all repeatedly accessing the same array from a readonly field. Changing them to instead access that same array but from a cached local results in an ~10% throughput boost on a microbenchmark that does uncontended reads/writes of objects from/to the queue. (I also tried using a ref local to point directly to the target slot in the array, but that actually resulted in a measurable regression.)
Andy Ayers [Tue, 27 Nov 2018 08:36:35 +0000 (00:36 -0800)]
Improve perf for Index based span indexers (#21196)
First, evaluate the actual index before invoking the underlying int indexer,
so that the jit doesn't think the span is address taken.
Second, improve the jit's ability to remove some redundant comparisons like
those that arise in computing the actual index.
Chuck Ries [Tue, 27 Nov 2018 06:06:34 +0000 (22:06 -0800)]
Fix a use after free for Managed Data BP (#21205)
ShimProxyCallback::DataBreakpoint::DataBreakpointEvent was holding onto a bare
BYTE* for the CONTEXT rather than copying the buffer and taking ownership. Due to
lifetime, this resulted in a use after free. Apparently in retail code we got lucky
and this worked enough of the time that we never noticed it.
Levi Broderick [Tue, 27 Nov 2018 05:08:10 +0000 (21:08 -0800)]
Fix Rune.ToUpper / ToLower under GlobalizationMode.Invariant (#21203)
Jan Vorlicek [Tue, 27 Nov 2018 03:15:45 +0000 (04:15 +0100)]
Disable noncollectible alc finalization (#21189)
* Disable noncollectible AssemblyLoadContext finalization
The constructor of AssemblyLoadContext was missing a call to
GC.SuppressFinalize for non-collectible AssemblyLoadContext. That means
that the finalizer can be executed and unloading being initialized even
for non-collectible AssemblyLoadContext.
This change fixes that.
* SuppressFinalize non-collectible AssemblyLoadContext
The finalizer for the non-collectible AssemblyLoadContext is never
called and so it should not be in the finalization queue.
* Handle finalizer call without constructor executed
Check if the constructor was executed before we run the finalizer code.
Jan Kotas [Tue, 27 Nov 2018 01:55:10 +0000 (17:55 -0800)]
Disable domain neutral loading (#21156)
Tanner Gooding [Tue, 27 Nov 2018 01:54:36 +0000 (17:54 -0800)]
Deleting a format.patch file that got checked in (#21202)
Jan Kotas [Tue, 27 Nov 2018 01:06:35 +0000 (17:06 -0800)]
Update documentation for changing CoreLib (#21201)
Brian Sullivan [Tue, 27 Nov 2018 00:31:37 +0000 (16:31 -0800)]
Merge pull request #21184 from mikedn/load-norm
Fix normalize on load handling in assertion propagation
Morgan Brown [Mon, 26 Nov 2018 22:46:36 +0000 (14:46 -0800)]
Fix Azure DevOps endpoint (#21038)
* Delete incorrect resources section
Remove reference to arcade repo
Tanner Gooding [Mon, 26 Nov 2018 21:27:46 +0000 (13:27 -0800)]
Updating runtest.py to support printing coredump information (#21149)
* Updating runtest.py to support printing coredump information
* Adding some comments to the coredump collection methods in runtest.py
Tanner Gooding [Mon, 26 Nov 2018 19:44:51 +0000 (11:44 -0800)]
Fixing a few small issues with the SIMD vs SIMD HWIntrinsics (#21097)
* Fixing the simdHandleCache to add entries for Vector64FloatHandle and Vector128FloatHandle
* Fixing two places that were calling just OperIsSIMD() rather than OperIsSIMDorSimdHWintrinsic()
Matt Mitchell [Mon, 26 Nov 2018 19:38:31 +0000 (11:38 -0800)]
Merge pull request #21194 from mmitche/fix-debug-checked
Don't attempt to sign xstate in debug/checked builds
Stephen Toub [Mon, 26 Nov 2018 19:08:38 +0000 (14:08 -0500)]
Remove TypedReference.MakeTypedReference readonly restriction (#21193)
TypedReference.MakeTypedReference validates that the fields aren't readonly, which limits our ability to make fields readonly in core while maintaining BinaryFormatter compatibility. This restriction isn't useful for a variety of reasons, though: reflection allows readonly fields to be set, but more importantly, TypedReference.SetTypedReference has always thrown NotSupportedException, so it's not even actually usable.
This change just removes the readonly check from MakeTypedReference.
Carol Eidt [Mon, 26 Nov 2018 19:04:56 +0000 (11:04 -0800)]
Merge pull request #21164 from kvochko/no_hw_intrinsics_x86
[x86] Do not support hardware intrinsics on x86 unix
Tanner Gooding [Mon, 26 Nov 2018 19:01:04 +0000 (11:01 -0800)]
Handle the remaining known issues (modulo snan) with the double/single parsing logic (#21036)
* Fixing the double/single parser to return negative zero for `-0`
* Updating the Utf8Parser to allow exponents of arbitrary length
* Updating the double/single parser to support case-insensitive and signed infinity/nan
* Ensure TryParseAsSpecialFloatingPoint handles the case where we have an empty source
* Fixing the number parser to allow `-0` for decimal and to not parse '++infini'
Michelle McDaniel [Mon, 26 Nov 2018 16:31:30 +0000 (08:31 -0800)]
Tar arm build outputs to workaround unicode issues (#21127)
Matt Mitchell [Mon, 26 Nov 2018 16:18:03 +0000 (08:18 -0800)]
Don't attempt to sign xstate in debug/checked builds
Mike Danes [Sun, 25 Nov 2018 09:22:08 +0000 (11:22 +0200)]
Don't generate copy assertions for normalize on load variables
Mike Danes [Sun, 25 Nov 2018 08:44:31 +0000 (10:44 +0200)]
Add test for 18867
Jeremy Koritzinsky [Sun, 25 Nov 2018 00:35:17 +0000 (18:35 -0600)]
Fix Interop/PInvoke/Miscellaneous/HandleRef tests under GCStress (#21131)
* Add GC.KeepAlive call in the IL stub when marshalling a HandleRef.
* Add a GC.KeepAlive call for the BoxedInt since it isn't kept alive when running GC Stress
* Reference HandleRef._handle via the binder instead of by field offset arithmetic
* Alignment
* Use Marshal.Alloc/FreeHGlobal instead of the weird no-pinning machinery that was failing before
* Update HandleRefTest.cs
Ayende Rahien [Sat, 24 Nov 2018 15:18:36 +0000 (17:18 +0200)]
Fixing typo (#21081)
Jan Kotas [Sat, 24 Nov 2018 15:00:36 +0000 (07:00 -0800)]
Refactor AppDomain.IsAppXModel and a few other AppDomain methods (#21181)
Contributes to #21028
Jan Kotas [Sat, 24 Nov 2018 14:59:54 +0000 (06:59 -0800)]
Move AppDomain local data store to AppContext and cleanup AppContext (#21180)
Contributes to #21028
Eugene Rozenfeld [Fri, 23 Nov 2018 22:55:02 +0000 (14:55 -0800)]
Don't assume objects don't escape in pure helpers.
We can't assume objects don't escape in helpers marked as pure for the following reasons:
1. The helpers may call user code that may make objects escape, e.g.,
https://github.com/dotnet/coreclr/blob/
c94d8e68222d931d4bb1c4eb9a52b4b056e85f12/src/vm/jithelpers.cpp#L2371
2. The helpers usually protect gc pointers with GCPROTECT_BEGIN() so the pointers are reported as normal pointers to the gc.
Pointers to stack-allocated objects need to be reported as interior so they have to be protected with
GCPROTECT_BEGININTERIOR().
3. The helpers may cause these asserts in ValidateInner on stack-allocated objects:
https://github.com/dotnet/coreclr/blob/
c94d8e68222d931d4bb1c4eb9a52b4b056e85f12/src/vm/object.cpp#L723
https://github.com/dotnet/coreclr/blob/
c94d8e68222d931d4bb1c4eb9a52b4b056e85f12/src/vm/object.cpp#L730
Jan Vorlicek [Fri, 23 Nov 2018 18:14:03 +0000 (19:14 +0100)]
Fix debugger collectible thread statics access (#21175)
The CordbClass::GetStaticFieldValue2 was missing support for accessing
thread statics in collectible classes. Fortunately the fix was simple,
we can use the same code path as for non-collectible thread statics.
dotnet-maestro-bot [Fri, 23 Nov 2018 17:34:42 +0000 (09:34 -0800)]
Update CoreClr, CoreFx, CoreSetup, PgoData to preview-27123-02, preview.18573.2, preview-27123-01, master-
20181123-0042, respectively (#21158)
John Doe [Thu, 22 Nov 2018 22:23:46 +0000 (14:23 -0800)]
Typos (#21171)
* accesible -> accessible
* acheive -> achieve
* actutl -> actual
* appied -> applied
* apropriate -> appropriate
* aprox -> approx
* aquired -> acquired
* argumemts -> arguments
* Argumens -> Arguments
* argumetn -> argument
Jan Kotas [Thu, 22 Nov 2018 22:22:40 +0000 (14:22 -0800)]
Delete System.AppDomainSetup (#21157)
* Delete System.AppDomainSetup
Contributes to #21028
* Add test hook for null entry assembly
* Validate that the binder paths are absolute
Michal Strehovský [Thu, 22 Nov 2018 18:09:37 +0000 (19:09 +0100)]
Mark the String.Empty field as Intrinsic (#21162)
This is required so that the CPAOT compiler can recognize this field as a RyuJIT intrinsic.
Michal Strehovský [Thu, 22 Nov 2018 17:38:16 +0000 (18:38 +0100)]
Narrow the assert in fgLateDevirtualization (#21163)
As suggested by Andy in dotnet/corert#6601.
dotnet-maestro[bot] [Thu, 22 Nov 2018 14:43:20 +0000 (14:43 +0000)]
Update dependencies from https://github.com/dotnet/arcade build 434 (#21166)
This change updates the following dependencies
- Microsoft.DotNet.Arcade.Sdk - 1.0.0-beta.18571.7
- Microsoft.DotNet.Helix.Sdk - 1.0.0-beta.18571.7
Andrey Kvochko [Thu, 22 Nov 2018 12:27:01 +0000 (15:27 +0300)]
Do not support HW intrinsics on x86 linux
Swaroop Sridhar [Thu, 22 Nov 2018 09:24:08 +0000 (01:24 -0800)]
Sequel to NativeLibrary API (#21148)
Small changes to address furhter feedback:
- Perform argument validation in the Marshall.cs instead of DllImport helpers
- Improve doc comments.
dotnet-maestro-bot [Thu, 22 Nov 2018 00:51:45 +0000 (16:51 -0800)]
Update CoreClr, CoreFx, CoreSetup to preview-27121-03, preview.18571.3, preview-27121-03, respectively (#21142)
Andrew Au [Thu, 22 Nov 2018 00:17:06 +0000 (16:17 -0800)]
Single stepping out of the unsafe places (#21135)
Aaron Robinson [Wed, 21 Nov 2018 21:59:57 +0000 (13:59 -0800)]
Merge pull request #21146 from AaronRobinsonMSFT/comtoclrworker_av
COMToCLRWorker AV
Aaron Robinson [Wed, 21 Nov 2018 19:21:24 +0000 (11:21 -0800)]
Minor clean up in COMToCLRWorker
Swaroop Sridhar [Wed, 21 Nov 2018 19:42:01 +0000 (11:42 -0800)]
Introduce Marshall.LoadLibrary API (#20871)
Implement Native LoadLibrary API
This change commits the following changes:
1) Refactoring DllImport code to reuse LodLibrary by search for
pInvoke and LoadLibrary cases
2) Implement the new Native Library API in System.Runtime.Interop.Marshall
3) Add tests for the new APIs
Eugene Rozenfeld [Mon, 19 Nov 2018 21:25:56 +0000 (13:25 -0800)]
Disable object stack allocation verification under GCStress.
ObjectStackAllocationTests use GC.GetAllocatedBytesForCurrentThread to
verify object stack allocations. Under GCStress the vm may initiate additional
heap allocations in GCHeap::StressHeap (see the call to 'pGenGCHeap->allocate' below).
This change re-enables ObjectStackAllocationTests and updates then to not verify stack allocations under GCStress.
It's useful to run the tests even without the verification to catch crashes, gc asserts, etc.
```
if (Interlocked::Increment(&OneAtATime) == 0 &&
!TrackAllocations()) // Messing with object sizes can confuse the profiler (see ICorProfilerInfo::GetObjectSize)
{
StringObject* str;
// If the current string is used up
if (HndFetchHandle(m_StressObjs[m_CurStressObj]) == 0)
{
// Populate handles with strings
int i = m_CurStressObj;
while(HndFetchHandle(m_StressObjs[i]) == 0)
{
_ASSERTE(m_StressObjs[i] != 0);
unsigned strLen = (LARGE_OBJECT_SIZE - 32) / sizeof(WCHAR);
unsigned strSize = PtrAlign(StringObject::GetSize(strLen));
// update the cached type handle before allocating
SetTypeHandleOnThreadForAlloc(TypeHandle(g_pStringClass));
str = (StringObject*) pGenGCHeap->allocate (strSize, acontext);
if (str)
{
str->SetMethodTable (g_pStringClass);
str->SetStringLength (strLen);
HndAssignHandle(m_StressObjs[i], ObjectToOBJECTREF(str));
}
i = (i + 1) % NUM_HEAP_STRESS_OBJS;
if (i == m_CurStressObj) break;
}
// advance the current handle to the next string
m_CurStressObj = (m_CurStressObj + 1) % NUM_HEAP_STRESS_OBJS;
}
// Get the current string
str = (StringObject*) OBJECTREFToObject(HndFetchHandle(m_StressObjs[m_CurStressObj]));
if (str)
{
// Chop off the end of the string and form a new object out of it.
// This will 'free' an object at the begining of the heap, which will
// force data movement. Note that we can only do this so many times.
// before we have to move on to the next string.
unsigned sizeOfNewObj = (unsigned)Align(min_obj_size * 31);
if (str->GetStringLength() > sizeOfNewObj / sizeof(WCHAR))
{
unsigned sizeToNextObj = (unsigned)Align(size(str));
uint8_t* freeObj = ((uint8_t*) str) + sizeToNextObj - sizeOfNewObj;
pGenGCHeap->make_unused_array (freeObj, sizeOfNewObj);
str->SetStringLength(str->GetStringLength() - (sizeOfNewObj / sizeof(WCHAR)));
}
else
{
// Let the string itself become garbage.
// will be realloced next time around
HndAssignHandle(m_StressObjs[m_CurStressObj], 0);
}
}
}
Interlocked::Decrement(&OneAtATime);
```
Aaron Robinson [Wed, 21 Nov 2018 19:07:41 +0000 (11:07 -0800)]
Check for null on COM to CLR call if error occurs during thread instance get.
Jan Kotas [Wed, 21 Nov 2018 18:24:11 +0000 (10:24 -0800)]
Change System.Object to ClassInterface.None (#21137)
* Change System.Object to ClassInterface.None
Prevents .NET Framework legacy interfaces from leaking out.
Fixes https://github.com/dotnet/corefx/issues/33638
* Use AutoDispatch for System.Object
Jan Vorlicek [Wed, 21 Nov 2018 18:20:01 +0000 (19:20 +0100)]
Fix Unix ARM JIT_MemCpy and JIT_MemSet (#21141)
The functions were incorrectly using 4 byte loads to probe for
the address validity. While the comment on JIT_MemCpy requires
4 byte aligned address, it doesn't match the way JIT uses it and
the Windows version of the function works with unaligned addresses
too.
This bug was discovered as a crash in an application where the
JIT_MemCpy was called with count=2 and an address that was two
bytes below the end of a memory page where the following page
was not mapped.
Tanner Gooding [Wed, 21 Nov 2018 16:37:35 +0000 (08:37 -0800)]
Ensuring that gc_stress and gc_stress_c are marked global when writing, but not when reading (#21128)
Jan Kotas [Wed, 21 Nov 2018 16:23:30 +0000 (08:23 -0800)]
Delete dead/unused code (#21138)
dotnet-maestro[bot] [Wed, 21 Nov 2018 14:43:28 +0000 (14:43 +0000)]
Update dependencies from https://github.com/dotnet/arcade build 420 (#21140)
This change updates the following dependencies
- Microsoft.DotNet.Arcade.Sdk - 1.0.0-beta.18570.17
- Microsoft.DotNet.Helix.Sdk - 1.0.0-beta.18570.17
dotnet-maestro-bot [Wed, 21 Nov 2018 12:20:33 +0000 (04:20 -0800)]
Update CoreSetup, PgoData to preview-27121-02, master-
20181121-0042, respectively (#21136)
dotnet-maestro-bot [Wed, 21 Nov 2018 07:26:39 +0000 (23:26 -0800)]
Update BuildTools, CoreClr, CoreFx, CoreSetup, IbcData, PgoData to preview1-03420-02, preview-27121-02, preview.18571.1, preview-27120-04, master-
20181120-0045, master-
20181120-0044, respectively (#21045)
Matt Mitchell [Wed, 21 Nov 2018 02:13:59 +0000 (18:13 -0800)]
Merge pull request #21129 from mmitche/sign-xstat
Sign api-ms-win-core-xstate-l2-1-0.dll
Stephen Toub [Wed, 21 Nov 2018 00:52:08 +0000 (19:52 -0500)]
Expose generic UnsafeQueueUserWorkItem (#21125)
Ben Adams [Wed, 21 Nov 2018 00:50:26 +0000 (00:50 +0000)]
Vectorize string.IndexOf(..., StringComparison.Ordinal) (#21076)
* Vectorize string.IndexOf(..., StringComparison.Ordinal)
* Merge two IndexOf(..., int* matchLengthPtr)
* fix
2 callers expect value.Length == 0 to be tested first
3rd caller ReplaceCore doesn't allow a value.Length == 0 so moving it
first is ok
Ben Adams [Wed, 21 Nov 2018 00:44:56 +0000 (00:44 +0000)]
Vectorize Array.{Last}IndexOf<T>(...) for byte and char (#21116)
* Vectorize Array.IndexOf<T>(...) for byte and char
* Also LastIndexOf
Carol Eidt [Tue, 20 Nov 2018 23:58:41 +0000 (15:58 -0800)]
Merge pull request #21058 from CarolEidt/FixVZeroUpper
Record AVX use for SIMD indirections
Matt Mitchell [Mon, 13 Aug 2018 18:15:05 +0000 (11:15 -0700)]
Sign api-ms-win-core-xstate-l2-1-0.dll
This file is only catalog signed in RS4. Sign during the build. To achieve this copy the CRT binaries locally rather than referencing from the UCRT location directly.
Stephen Toub [Tue, 20 Nov 2018 22:03:33 +0000 (17:03 -0500)]
Remove remaining StringBuilder marshaling use from Corelib (#21120)
* Remove remaining StringBuilder marshaling use from Corelib
- Unix globalization functions were using StringBuilder. Replaced them with stackallocs, as the max sizes were all relatively small.
- Registry methods were declared to use StringBuilder, but these weren't actually used by corelib, and in fact, corefx isn't using StringBuilder with them either. I've changed the definitions for now to use char[] instead (all the call sites are passing in null), and I'll fix up some corefx usage separately.
- Resource-related functions were using StringBuilder, and have been replaced with stackallocs, given reasonably small max buffer sizes.
- ExpandEnvironmentVariables was using a StringBuilder, but it's rewritten equivalent code in corefx is not. For now I've copied the corefx function over to coreclr to replace the different implementation, but we can look at subsequently consolidating them into one.
* Address PR feedback
Stephen Toub [Tue, 20 Nov 2018 20:59:37 +0000 (15:59 -0500)]
Remove StringBuilder from HebrewNumber formatting (#21122)
Rather than allocating a new StringBuilder, writing into that, getting its string, and then appending that string to another StringBuilder, we can just write directly to the original one.
Eugene Rozenfeld [Tue, 20 Nov 2018 01:08:06 +0000 (17:08 -0800)]
Fix for #21011: propagate GTF_DONT_CSE on comma returns
When a method returns a multi-reg struct, we set GTF_DONT_CSE on return local: https://github.com/dotnet/coreclr/blob/
497419bf8f19c649d821295da7e225e55581cce9/src/jit/importer.cpp#L8783
Setting GTF_DONT_CSE blocks assertion propagation here:
https://github.com/dotnet/coreclr/blob/
9d49bf1ec6f102b89e5c2885e8f9d3d77f2ec144/src/jit/assertionprop.cpp#L2845-L2848
In the test we have a synchronized method so we change the return node to
return a comma that include a call to HELPER.CORINFO_HELP_MON_EXIT.
If the rightmost comma expression doesn't have GTF_DONT_CSE,
assertion propagation is not blocked and we end up with this tree:
```
[000040] -----+------ /--* CNS_INT struct 0
[000043] --C-G+------ /--* COMMA struct
[000036] --C-G+------ | \--* CALL help void HELPER.CORINFO_HELP_MON_EXIT
[000032] L----+------ arg1 in x1 | +--* ADDR long
[000031] ----G+-N---- | | \--* LCL_VAR ubyte (AX) V03 tmp1
[000033] -----+------ arg0 in x0 | \--* LCL_VAR ref V00 this
[000041] -AC-G+------ * COMMA struct
[000006] -----+-N---- | /--* LCL_VAR struct V01 loc0
[000039] -A---+------ \--* ASG struct (copy)
[000037] D----+-N---- \--* LCL_VAR struct V05 tmp3
```
Downstream phases can't handle struct zero return expressed as
```
[000040] -----+------ /--* CNS_INT struct 0
```
The fix is to propagate GTF_DONT_CSE to the rightmost comma expression
to block bad assertion propagation.
Fixes #21011.
Tanner Gooding [Tue, 20 Nov 2018 20:01:25 +0000 (12:01 -0800)]
Fixing up some issues with the HWIntrinsic test templates (#21104)
* Updating the VectorNotSupportedTest to throw on failure
* Updating the HWIntrinsic test templates to check success on a per-scenario basis
* Regenerating the templated HWIntrinsic tests
Stephen Toub [Tue, 20 Nov 2018 19:58:47 +0000 (14:58 -0500)]
Replace custom AnsiToLower with more efficient ToLowerAsciiInvariant (#21121)
Avoids all of the StringBuilder-related costs.
Jan Kotas [Tue, 20 Nov 2018 18:56:12 +0000 (10:56 -0800)]
Delete CORCOMPILE_IMPORT_TABLE_ENTRY (#21109)
Tom Deseyn [Tue, 20 Nov 2018 18:52:46 +0000 (19:52 +0100)]
Linux: FlushProcessWriteBuffers using membarrier when available (#20949)
* Linux: FlushProcessWriteBuffers using membarrier when available
* Ensure __NR_membarrier is defined for x64 portable build.
* Define __NR_membarrier on all portable architectures
Tanner Gooding [Tue, 20 Nov 2018 15:56:52 +0000 (07:56 -0800)]
Adding some AggressiveInlining attributes to the Vector64/128/256 helper methods (#21111)
dotnet-maestro[bot] [Tue, 20 Nov 2018 14:53:47 +0000 (14:53 +0000)]
Update dependencies from https://github.com/dotnet/arcade build 402 (#21118)
This change updates the following dependencies
- Microsoft.DotNet.Arcade.Sdk - 1.0.0-beta.18569.4
- Microsoft.DotNet.Helix.Sdk - 1.0.0-beta.18569.4
Stephen Toub [Tue, 20 Nov 2018 10:08:00 +0000 (05:08 -0500)]
Improve async iterator exception stack traces (#21103)
- The current stack trace prettying logic assumes that a method only has one StateMachineAttribute, but that's no longer true with async iterators, which (at least with the current implementation) can have both an AsyncStateMachineAttribute and an IteratorStateMachineAttribute. As such, we change the prettying logic to look through all StateMachineAttributes rather than stopping at the first one.
- ManualResetValueTaskSourceCore.GetResult is responsible for propagating exceptions, but it's infrastructure and isn't particularly meaningful in a call stack. Hide it from the stack trace, as we do for other such GetResult methods in Task and ValueTask.
Sergey Andreenko [Tue, 20 Nov 2018 05:53:14 +0000 (21:53 -0800)]
Make code in `getTargetMethodDesc` more reliable. (#21101)
* Make code in `getTargetMethodDesc` more reliable.
Fixes DevDiv_723667.
* change the condition to make sure that we do not have false true cases.
Sergey Andreenko [Tue, 20 Nov 2018 05:39:49 +0000 (21:39 -0800)]
format spmi sources (#21094)
We do not do it automatically because we do not change it often.
Hugh Bellamy [Tue, 20 Nov 2018 05:31:30 +0000 (05:31 +0000)]
Cleanup System.Reflection.Emit code and delete dead code (#21086)
* Cleanup System.Reflection.Emit code
* Delete dead code from System.Reflection.Emit
Jan Kotas [Tue, 20 Nov 2018 05:30:40 +0000 (21:30 -0800)]
Update xunit version to match corefx (#21098)
* Update xunit version to match corefx
Fixes https://github.com/dotnet/corefx/issues/33587
* Disabled outdated test
Carol Eidt [Tue, 20 Nov 2018 03:43:03 +0000 (19:43 -0800)]
Add asserts for the types of the Span indexers. (#21037)
Also, add a test for #20958
Omair Majid [Tue, 20 Nov 2018 00:32:13 +0000 (19:32 -0500)]
Fix search for headers and libraries via LLVM_HOME (#21084)
Commit
b89e2305 (PR #18873) renamed WITH_LLDB_LIBS to LLDB_LIB_DIR and
WITH_LLDB_INCLUDES to LLDB_INCLUDE_DIR. But it forgot to update the
usage of these variables in some find_library and find_path calls.
As a result, setting LLVM_HOME is no longer enough to compile coreclr on
platforms with non-standard locations for lldb. This commit fixes that.
Tarek Mahmoud Sayed [Mon, 19 Nov 2018 20:50:33 +0000 (12:50 -0800)]
Fix Getting Max Japanese Era (#21087)
We were using the ICU API ucal_getLimit and askin gto get the maximum value of the Japanese eras. it looks this API is just return the era matching the current system clock which prevent returning any future defined era in the system (which ICU 63.1 can support with some environment variable).
We raised the issue to the ICU owners but would be better just to get the era of Gregorian year 9999 which always should return the max era anyway.
Stephen Toub [Mon, 19 Nov 2018 20:45:25 +0000 (15:45 -0500)]
Disable stale sockets test (#21083)
Hugh Bellamy [Sat, 17 Nov 2018 19:05:39 +0000 (19:05 +0000)]
Cleanup RtType.cs code
Hugh Bellamy [Sat, 17 Nov 2018 18:45:30 +0000 (18:45 +0000)]
Fix Type.GetInterface for null namespace and non-null interface namespace
dotnet-maestro[bot] [Mon, 19 Nov 2018 15:23:43 +0000 (15:23 +0000)]
Update dependencies from https://github.com/dotnet/arcade build 385 (#21082)
This change updates the following dependencies
- Microsoft.DotNet.Arcade.Sdk - 1.0.0-beta.18568.1
- Microsoft.DotNet.Helix.Sdk - 1.0.0-beta.18568.1
Jacek Blaszczynski [Mon, 19 Nov 2018 11:29:11 +0000 (12:29 +0100)]
Add headers to crossgen, cee_crossgen, corzap_crossgen vcxproj files (#20082)
Michal Strehovský [Mon, 19 Nov 2018 08:23:22 +0000 (09:23 +0100)]
Make type comparisons more general purpose (#20940)
This has two parts:
## Part 1
CoreRT represents native type handles differently from CoreCLR - on CoreCLR, `RuntimeTypeHandle` is a wrapper over `RuntimeType` and RyuJIT is aware of that. On CoreRT, `RuntimeTypeHandle` wraps the native type handle, not a `RuntimeType`.
The knowledge is hardcoded in importer when importing the sequence "ldtoken foo / call Type.GetTypeFromHandle" - importer just removes the call and bashes the result of ldtoken to be a reference type. CoreRT had to avoid reporting `Type.GetTypeFromHandle` as an intrinsic because of that.
I'm adding another helper that lets RyuJIT avoid hardcoding that knowledge. Instead of just bashing the return type, we swap the helper call.
## Part 2
Native type handle equality checks need to go through a helper, unless the EE side says it's okay to compare native type handles directly.
Ben Adams [Mon, 19 Nov 2018 06:18:57 +0000 (06:18 +0000)]
Single pass Vectorize Span<byte>.IndexOfAny(..., ReadOnlySpan<byte>) for 2,3 lengths (#20738)
Jacek Blaszczynski [Sun, 18 Nov 2018 20:08:31 +0000 (21:08 +0100)]
Add headers to all md*, v3binder*, and ceefgen vcxproj (#20163)
Work toward #14884
Hugh Bellamy [Sun, 18 Nov 2018 18:31:28 +0000 (18:31 +0000)]
Add ICustomMarshaler tests (#19195)
* Add ICustomMarshaler tests
dotnet-maestro[bot] [Sun, 18 Nov 2018 14:50:14 +0000 (14:50 +0000)]
Update dependency files (#21067)
* Update dependencies from https://github.com/dotnet/arcade build 379
This change updates the following dependencies
- Microsoft.DotNet.Arcade.Sdk - 1.0.0-beta.18566.6
- Microsoft.DotNet.Helix.Sdk - 1.0.0-beta.18566.6
* Update dependencies from https://github.com/dotnet/arcade build 382
This change updates the following dependencies
- Microsoft.DotNet.Arcade.Sdk - 1.0.0-beta.18567.1
- Microsoft.DotNet.Helix.Sdk - 1.0.0-beta.18567.1
Hugh Bellamy [Sun, 18 Nov 2018 04:08:27 +0000 (04:08 +0000)]
Fix incorrect exception thrown in Marshal.ReleaseComObject(null) (#21072)
Tanner Gooding [Sat, 17 Nov 2018 18:24:58 +0000 (10:24 -0800)]
Removing the legacy helper intrinsics and adding tests for their replacements (#20994)
* Removing helper intrinsics from the x86 APIs in S.P.Corelib
* Removing JIT support for the removed x86 helper intrinsics
* Removing the x86 HardwareIntrinsics tests for the removed helper APIs
* Fixing up existing usages to no longer use the removed x86 helper intrinsics
* Skip CoreFX tests dependent on the removed x86 helper intrinsics
* Adding a GenerateTests.csx and templates for the new shared helper intrinsics
* Generating the new shared helper intrinsics from their templates
* Disabling some tests for arm and arm64 that are failing due to an assert
Egor Chesakov [Sat, 17 Nov 2018 18:20:13 +0000 (10:20 -0800)]
Force Unicode in test_location parameter in delete_existing_wrappers() (#21056)
Nick Craver [Sat, 17 Nov 2018 08:11:54 +0000 (03:11 -0500)]
Optimization: Further reduce allocations in GetCustomAtttributes() paths (#20795)
* Remove PseudoCustomAttribute allocations
Inside CustomAttribute, PseudoCustomAttribute is the mechanism used to return Attribute instances for what are really flags and such underneath in most cases. This is just an implentation detail, but this detail creates an intermediate array just for passing things around inside CustomAttribute...that's not necessary. This replaces the Attribute[] allocations with the ListBuilder<Attribute> struct for accumulating and relaying this attributes to parent methods.
Many overloads had a lot of duplicate coalescing code on this path as well...DRYed up that series of methods.
This change reduces allocation overhead for all calls by 40 bytes. In the case of a miss where none would be returned (e.g. .GetCustomAttributes<Attribute>() on a non-attributed class), this reduces allocations down to zero.
* CustomAttribute: split GetCustomAttributes into total and additive use cases
The split here was simple, but allocated an extra intermediate array for each parent type when searching for attributes in the inheritance case. Instead of doing that, we now move that to the second use case exclusively (which uses/returns the array created). The other use case adds to the ListBuilder<object> struct to accumulate through the inheritance hierarchy crawl (rather than an after-the-fact copy).
Note some parameter removals on the filter side...those weren't used already (and got in the way of this change).
* CustomAttributes: Cleanup of unused ops
Just removing the cruft that isn't actually used in here anymore (some of it in quite some time). Trivial amounts of speed gain, but every little bit helps!
* Param typo and red passing for GetCombinedList()
* Assume zero pseudo attributes by default (the most likely case)
Thsi eats the resize on the > 1 case, but that's so rare it's a net win to start with 0 here.
* Pseudo attributes: use out instead of return
A bit if a performance boost from the struct copy removal
* Pseudo attributes: switch to out everywhere and remove uneeded code paths
In relation to #20795, these old calls seem to be for consistency but they're not likely to change and a comment (I think) accomplishes the same goal with less code...so here's a commit demonstrating it.
* Comment spacing for pseudo attribute fetching
Make it clear that pseudo attributes are a distinct thing (e.g. why not in *this* method) than the return line.
Sergey Andreenko [Sat, 17 Nov 2018 01:38:00 +0000 (17:38 -0800)]
Exclude GC stress failure test. (#21059)
Sergey Andreenko [Sat, 17 Nov 2018 00:51:12 +0000 (16:51 -0800)]
Exclude tests from GCStress failing due to 21044 (#21060)
Bruce Forstall [Sat, 17 Nov 2018 00:42:58 +0000 (16:42 -0800)]
Merge pull request #21004 from BruceForstall/UpdateRuntestHelp
Update help output of runtest.cmd/sh
Zeng Jiang [Sat, 17 Nov 2018 00:25:43 +0000 (08:25 +0800)]
Add PInvoke/Decimal tests (#19301)
* Add PInvoke/Decimal Tests
* Remove tests that fail on both .NET Framework and .NET Core.
* Refactor away the Windows-only headers.
* Make Decimal tests build on Linux.
* Remove some dead code.
* Clean up formatting.
* Don't run Decimal <--> Currency tests off-Windows.
* Make calling conventions and parameter/return marshalling more explicit
Jan Kotas [Fri, 16 Nov 2018 21:00:41 +0000 (13:00 -0800)]
Delete unnecessary StackCrawlMarks in RtFieldInfo (#21054)
Left over from Code Access Security
Fixes #20965
Jan Vorlicek [Fri, 16 Nov 2018 18:08:29 +0000 (19:08 +0100)]
Fix unloadability races (#21024)
* Fix LoaderAllocator::AllocateHandle
When another thread wins the race in growing the handle table, the code
was not refreshing the slotsUsed local to the new up to date value. This
was leading to overwriting / reusing a live handle.
This change fixes it.
* Embed ThreadLocalBlock in Thread
Instead of allocating ThreadLocalBlock dynamically, embed it in the
Thread. That solves race issue between thread destruction and
LoaderAllocator destruction. The ThreadLocalBlock could have been
deleted during Thread shutdown while the LoaderAllocator's destruction
would be working with it.
Tarek Mahmoud Sayed [Fri, 16 Nov 2018 17:02:02 +0000 (09:02 -0800)]
Fix Formatting Japanese First Year on Linux (#21039)
When formatting the first year with Gannen symbol we are checking if the year is followed by single quote which this is the case on Windows but it is not neccessay the case on Linux
Carol Eidt [Fri, 16 Nov 2018 17:01:57 +0000 (09:01 -0800)]
Record AVX use for SIMD indirections
Jan Kotas [Fri, 16 Nov 2018 07:11:28 +0000 (23:11 -0800)]
Add workaround for https://github.com/dotnet/coreclr/issues/21051
Jan Kotas [Fri, 16 Nov 2018 06:33:52 +0000 (22:33 -0800)]
Add explicit signatures for Span indexer
Fixes #21032 and #20958
Dan Moseley [Fri, 16 Nov 2018 15:14:11 +0000 (07:14 -0800)]
Eliminate CMAKE dependency when not required (#21040)
dotnet-maestro[bot] [Fri, 16 Nov 2018 14:53:21 +0000 (14:53 +0000)]
Update dependencies from https://github.com/dotnet/arcade build 365 (#21052)
This change updates the following dependencies
- Microsoft.DotNet.Arcade.Sdk - 1.0.0-beta.18565.9
- Microsoft.DotNet.Helix.Sdk - 1.0.0-beta.18565.9
Jeremy Koritzinsky [Fri, 16 Nov 2018 07:48:16 +0000 (23:48 -0800)]
Add simple tests for ArrayWithOffset marshaling. (#21013)
* Add simple tests for ArrayWithOffset marshaling.
* PR Feedback.
* Fix visibility
* Fix DllImport library name
Zeng Jiang [Fri, 16 Nov 2018 07:47:51 +0000 (15:47 +0800)]
Add Interop/PInvoke/Attributes tests (#19107)
* Add Interop/PInvoke/Attributes tests
* Fix PR comments
* Fix compile warnings
* Remove comment
* Clean up test files to match current build system.
* Fix up test and make it windows only (as FEATURE_LCID is only enabled on Windows)
* Fixed string const issues.
Zeng Jiang [Fri, 16 Nov 2018 07:46:51 +0000 (15:46 +0800)]
Add PInvoke/CriticalHandles tests (#19297)
* Add PInvoke/CreticalHandles tests
* Bring up to current infrastructure.
* Fix ReverseTest.